Latency: will it undermine the most interesting 5G use cases?
One of the big topics at the OpenStack conference was how to prepare and implement a successful transition to 5G. Unfortunately, this means that the age-old Telco bugbears will rear their ugly heads again, including latency. 5G, as a fundamental requirement, mandates a 1 millisecond latency from the datasource to its destination. Currently, most applications are hosted in a small number of data centers (1-5), each of which is in charge of a geographic region. These have to communicate with each other.
While current network speeds may be enough to meet the latency requirements of 4G applications, 5G will necessitate a change, if only because the continental US is ~60ms wide, meaning that a datacenter on one coast communicating with another datacenter on the opposite coast will be too slow for 5G. With the 5G revolution, operators will need to manage hundreds of edge deployments, and maintain the physical space and hardware to achieve 1ms of latency. For some, this signals the need to tune at the bare metal level, leading to the one of the great ironies at the OpenStack conference: half of those involved with the project talked about increased virtualization, and half talked about de-virtualizing virtualization.
The spotlight is on 5G due a number of alluring use cases, including autonomous surgery, drone control, and self-driving/remote control cars. As appealing as these may sound, they are unlikely to be realized in the near term. As an example, many proposed implementations of self-driving capabilities, such as Tesla’s, only check in to a datacenter to update maps and report different road conditions than expected, neither of which needs to happen in real time, which somewhat undermines the 5G use case. Autonomous surgery also is somewhat unlikely, given the potentially tricky legal and medical complications and requirements.
If these high-visibility, big-ticket applications are not good 5G use cases, what would make the transition costs worth the investment?
In fact, 5G has plenty of valid use cases, one of which is virtual reality. If I am wearing a VR headset and turn my head, the network has to send the information for what to display before I would see it. This requires 1 ms network latency. Successful implementation of a VR service could have massive implications for a number of industries, such as retail. Imagine virtual retail environments, where customers browse virtual shelves and products, and converse with customers who share interests and needs. These types of applications justify the cost of 5G.
Here at VoltDB, we’re excited to see what else 5G will bring. VoltDB is one of the only databases in the world to be proven by Telcos, and customers in other industries, to successfully manage high-performance transactions in the millisecond range, quickly and correctly. For many realistic, market-driven use cases, VoltDB is the best way to migrate to 5G.